Generation of UAV-based Training Dataset using Semi-supervised Learning

Authors

DOI:

https://doi.org/10.18372/1990-5548.72.16935

Keywords:

dataset formation, semi-supervised learning, pseudo-labeling, unmanned aerial vehicle, YOLOv5, object detection, classification problem

Abstract

The paper considers the problem of constructing a training sample based on the use of semi-supervised learning a teacher. The problem statement related to the problem posed is substantiated. It is shown that obtaining a training sample in some cases is a difficult task that requires significant computational and financial costs. The use of semi-supervised learning made it possible to label unlabeled data and thus ensure the creation of a labeled sample of sufficient size. The paper gives examples of generating a training sample, as well as its use for training neural networks, which are used to solve the problem of multiclass classification. Using this approach, you can get a robust data set consisting of a small amount of manually labeled images and a huge amount of pseudo-labeled or augmented data. Using this approach, one can train a classifier to detect and classify any objects in images with bounding boxes and label them accordingly.

Author Biographies

Victor Sineglazov , National Aviation University, Kyiv

Doctor of Engineering Science. Professor. Head of the Department.

Aviation Computer-Integrated Complexes Department

Faculty of Air Navigation Electronics and Telecommunications

Vadym Kalmykov , National Aviation University, Kyiv

Post-graduate student

Faculty of Air Navigation, Electronics and Telecommunications

References

V. M. Sineglazov and V. V. Kalmykov, “Image Processing from Unmanned Aerial Vehicle Using Modified YOLO Detector,” Electronics and control systems, NAU Kyiv: vol. 3, no. 69, pp.37–42, 2021. https://doi.org/10.18372/1990-5548.69.16425

J. H. Kim, J. Kim, S. J. Oh, S. Yun, H. Song, J. Jeong, & H. O. Song, (2022). Dataset Condensation via Efficient Synthetic-Data Parameterization. arXiv preprint arXiv:2205.14959.

M. Maranghi, A. Anagnostopoulos, I. Cannistraci, I. Chatzigiannakis, F. Croce, G. Di Teodoro, & P. Velardi, (2022). AI-based Data Preparation and Data Analytics in Healthcare: The Case of Diabetes. arXiv preprint arXiv:2206.06182.

C. Shneider, A. Hu, A. K. Tiwari, M. G. Bobra, K. Battams, J. Teunissen, & E. Camporeale, (2021). A Machine-Learning-Ready Dataset Prepared from the Solar and Heliospheric Observatory Mission. arXiv preprint arXiv:2108.06394.

Semi-supervised learning, 2022, https://en.wikipedia.org/wiki/Semi-supervised_learning

Ross Girshick, Jeff Donahue, Trevor Darrell, and Jitendra Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In CVPR 2014. https://doi.org/10.1109/CVPR.2014.81

Kaiming He, Georgia Gkioxari, Piotr Doll´ar, and Ross Gir-shick. Mask r-cnn. In ICCV 2017.

Zhaowei Cai and Nuno Vasconcelos. Cascade r-cnn: Delving into high quality object detection. In CVPR 2018.

Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. You only look once: Unified, real-time object detection. In CVPR 2016. https://doi.org/10.1109/CVPR.2016.91

Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian Szegedy, Scott Reed, Cheng-Yang Fu, and Alexander C Berg. Ssd: Single shot multibox detector. In ECCV 2016. https://doi.org/10.1007/978-3-319-46448-0_2

Jisoo Jeong, Seungeui Lee, Jeesoo Kim, and Nojun Kwak. Consistency-based semi-supervised learning for object detection. In NeurIPS, 2019.

Antti Tarvainen and Harri Valpola. Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. In NeurIPS, 2017.

Qizhe Xie, Zihang Dai, Eduard Hovy, Minh-Thang Luong, and Quoc V. Le. Unsupervised data augmentation for consistency training. arXiv preprint arXiv:1904.12848, 2019.

Daniel Ho, Eric Liang, Ion Stoica, Pieter Abbeel, and Xi Chen. Population based augmentation: Efficient learning of augmentation policy schedules. arXiv preprint arXiv:1905.05393, 2019.

Ekin D Cubuk, Barret Zoph, Dandelion Mane, Vijay Vasudevan, and Quoc V Le. Autoaugment: Learning augmentation strategies from data. In CVPR, 2019. https://doi.org/10.1109/CVPR.2019.00020

Aerial dataset. Website, 2022. https://universe.roboflow.com/gdit/aerial-airport/dataset/1.

Wang, C. Y., Liao, H. Y. M., Wu, Y. H., Chen, P. Y., Hsieh, J. W., & Yeh, I. H. (2020). CSPNet: A new backbone that can enhance learning capability of CNN. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops (pp. 390–391).

Li, Z., Peng, C., Yu, G., Zhang, X., Deng, Y., & Sun, J. (2017). Light-head r-cnn: In defense of two-stage object detector. arXiv preprint arXiv:1711.07264.

Downloads

Published

2022-09-23

Issue

Vol. 2 No. 72 (2022)

Section

COMPUTER SCIENCES AND INFORMATION TECHNOLOGIES

License

Authors who publish with this journal agree to the following terms:

Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.

Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.

Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).